A cuproptosis-related gene cluster in prediction of ovarian cancer prognosis and chemotherapeutic response

doi:10.21203/rs.3.rs-2320013/v1

Download PDF

Research Article

A cuproptosis-related gene cluster in prediction of ovarian cancer prognosis and chemotherapeutic response

https://doi.org/10.21203/rs.3.rs-2320013/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Ovarian cancer (OC) is a common gynecological malignancy correlated with a poor prognosis. Cuproptosis is a newly discovered form of cell death and has a close relationship with cancers, but the relationship between OC and cuproptosis remains unclear.

Results

In this study, we explored cuproptosis-related genes (CRGs) in public databases and found most CRGs are closely related to survival, and the potential roles of FDX1, LIAS and SLC31A1 in proliferation and migration were discovered in OC cell line. Afterwards, all 791 OC patients were divided into 2 clusters and the pathway enrichment and survival time showed obvious difference. 70 differentially expressed genes between 2 clusters were utilized to construct a gene signature. Significant difference was found in survival time and tumor-infiltrating immune cells among different risk groups. Finally, sensitivity of 12 commonly-used chemotherapeutic drugs was predicted closely correlating with risk score, which may provide a new strategy for clinical practice.

Conclusion

Cuproptosis

Ovarian cancer

cancer progression

prognosis

chemotherapeutic response

Ovarian cancer (OC) is a common gynecological malignancy worldwide and its incidence and mortality rate rank eighth among all malignant tumors in female. In 2020, over 310,000 women were newly diagnosed with OC and more than 200,000 patients died because of OC(1). Compared with other gynecological malignant tumors such as cervical cancer or endometrial cancer, despite significant progress in therapeutic drugs and comprehensive staging operation against OC, the overall survival (OS) of OC remains poor in most countries over the last several decades(2, 3). Therefore, finding new prediction and treatment targets for OC has become a top priority.

Copper is a crucial trace element which maintains a delicate balance in the human body, and the imbalance of copper in vivo can lead to serious physical disorders(4). The ATP7B mutation can cause Wilson’s disease due to abnormal copper accumulation(5), whereas the ATP7A mutation can cause Menke’s disease with cooper deficiency(6). Simultaneously, the level of copper is demonstrated to be anomalous in many tumors, including but not limited to breast, thyroid, colorectal, lung and oral cancers(7–12), while the therapeutic efficacy of copper complexes as anticancer modalities is also demonstrated by many clinical trials(13, 14).

Regulated cell death (RCD) is an autonomously and orderly cell death mode determined by genes and signaling molecules, which is ubiquitous in various biological processes(15). A growing number of RCD forms have been discovered involved in various physiological and pathological pathways, such as apoptosis, necroptosis, ferroptosis, pyroptosis, autophagy, alkaliptosis, lysosome-dependent cell death and more(16). Recently, a novel copper-dependent RCD mode was discovered by Tsvetkov et al.(17) named “cuproptosis”. In this study, copper can directly bind to lipoylated proteins of the pyruvate dehydrogenase complex under the regulation of FDX1 gene and affect the tricarboxylic acid (TCA) cycle and mitochondrial respiration so as to trigger Fe-S cluster proteins loss and ultimately cause cell death. Many studies have demonstrated the relevance between cuproptosis and cancer. Zhang et al. built and validated a novel cuproptosis-related lncRNA signature to predict prognosis in hepatocellular carcinoma(18). Han et al. also established a signature of lncRNA in soft tissue sarcoma and proved its liability(19). Lv et al. analyzed cuproptosis-related genes and explored the mechanism between LIPT1 and immune infiltration in melanoma(20). Bian et al. utilized cuproptosis-related genes to build a model for predicting prognosis in renal clear cell carcinoma(21). However, the role of cuproptosis in OC is still unclear.

In the present study, we explored cuproptosis-related genes (CRGs) in 791 OC patients from public databases and constructed a risk model for predicting OS, so as to provide a novel perspective for exploring the treatment of OC.

Expression of Cuproptosis-related genes in ovarian cancer

As mentioned in previous studies, we identified 19 cuproptosis-related genes (CRGs) including ATP7A, ATP7B, CDKN2A, DBT, DLAT, DLD, DLST, FDX1, GCSH, GLS, LIAS, LIPT1, LIPT2, MTF1, NFE2L2, NLRP3, PDHA1, PHDB and SLC31A1. To find out the relationship between CRGs and ovarian cancer, 88 cases of normal ovary and 379 cases of ovarian cancer were obtained from GTEx dataset and TCGA dataset, respectively. All those 19 genes showed a remarkable difference between cancer and normal tissues, while most of them were up-regulated in cancer and seven genes (NFE2L2, LIAS, LIPT1, PDHA1, GLS, GCSH, DLST) showed the opposite tendency (Fig. 1A). Some CRGs also showed a correlation with clinicopathological features, including tumor FIGO stage and grade in ovarian cancer tissues (Fig. 1B and C). We also discovered 5.63% of OC patients have CRGs mutations (Fig. 1D) while NLRP3 is the most frequent. Copy number variation (CNV) frequencies of CRGs in OC patients were depicted in Fig. 1E, and LIPT2 had the highest CNV gain. The position of CRGs on chromosomes and their CNV frequencies are shown in Fig. 1F. Besides, we investigated the correlation between the expression of CRGs, which revealed a mostly positive correlation of each other (Fig. 1G).

To better discover the function of CRGs in OC, we obtained the transcriptome data and clinical data from both TCGA dataset (376 cases) and GSE13876 of GEO dataset (415 cases). In total, 791 OC patients in all were analyzed after the combination of two datasets. As we can see in Fig. 2A, the expression of 12 CRGs was significantly related to the OS of OC, while higher expression of FDX1 and CDKN2A was associated with longer OS, and the other 10 CRGs showed opposite effect. Cox analysis revealed that CDKN2A and PDHA1 were favorable factors and the other 14 CRGs were risk factors for OC (Fig. 2B).

Crgs Correlate With Oc Proliferation And Migration In Vitro

To determine the potential role of CRGs in OC, we utilized small interfering RNAs (siRNAs) to knockdown the expression of several critical CRGs in OC cell line SK-OV-3, including FDX1, LIAS and SLC31A1. The efficiency of siRNAs was confirmed by qRT-PCR (Fig. 3A). As shown in Fig. 3B-D, we found that FDX1 knockdown could significantly increase the cell proliferation and clone formation ability, while LIAS and SLC31A1 knockdown showed the opposite effect. In addition, transwell and wound healing assays were performed and we found that FDX1 knockdown increased the migration ability of SK-OV-3 cells, but LIAS and SLC31A1 knockdown decreased cell migration ability (Fig. 3E-H).

Consensus Cluster Analysis Of Crgs Identified Two Oc Subgroups

To better understand the relationship between cuproptosis and ovarian cancer, consensus clustering analysis was adopted and the merged dataset was divided into two CRG-clusters (Fig. 4A-B). Compared with CRG-cluster A, most CRGs in CRG-cluster B were highly expressed (Fig. 4C), whereas patients in CRG-cluster B showed a shorter OS (Fig. 4D).

GSVA between these two clusters found enrichment in many pathways associated with tumorigenesis, such as translation initiation, telomere maintenance, mismatch repair, spliceosome, JAK-STAT signaling pathway and so on (Figure S1A-B). CIBERSORT algorithm was also performed to calculate the percentage of different immune cells and explore the relationship between cuproptosis and the tumor immune microenvironment. As can be seen in Figure S1C, the infiltration of activated B cell, immature B cell, eosinophil, mast cell, MDSC, macrophage, neutrophil, Natural Killer cell, Natural Killer T cell, CD8 positive T cell, regulatory T cell, Th1, Th2 and Th17 cell showed remarkably difference between two CRG clusters.

Genes Expressed Differently Between Two Crg-related Subgroups

Following that, we look for differentially-expressed gene among two groups, using a cut-off value of |Log2FC|>0.585 and p value < 0.01. In cluster B, 65 genes were significantly high expressed, while 5 genes were less expressed (Fig. 5A). GO pathway analysis was used and we noticed that those 70 differentially-expressed genes were mainly enriched in fatty acid metabolism and acetyl-CoA biosynthetic (Fig. 5B), which is consistent with previous studies. Then, we again used consensus cluster analysis to divide these patients into two gene-clusters based on the 70 genes above (Fig. 5C). The heatmap displayed the expression of those 70 genes in two gene-clusters and two CRG-clusters, and most of the 70 differentially-expressed genes were highly-expressed in gene-cluster A (Fig. 5D). 13 CRGs were significantly higher expressed in gene-cluster A (Fig. 5E), while patients included in gene-cluster A had a shorter OS (Fig. 5F).

Construction Of A Crgs-related Prognostic Signature In Oc

To further investigate the biological function of CRGs in OC, LASSO Cox regression analysis was performed to discover the relationship between 70 differentially-expressed genes and the OS of OC, and ultimately 10 genes were recognized (Fig. 6A-C). Next, multivariate Cox analysis identified 6 genes for calculating the risk score, and we identified the risk score as BMI1×0.174007772155564 + NRIP1×0.126565997146839 + PSAT1×0.0887611366137917 + MAD2L1×0.0978795316405013 + SCG5×0.0914323350379329 + CXCL14×0.116043990270974.

In Fig. 6D, we noticed that 12 CRGs (NFE2L2, ATP7A, ATP7B, SLC31A1, DLD, DLAT, LIAS, PDHB, MTF1, GLS, DBT and GCSH) were markedly overexpressed in the higher risk score group. Additionally, OC patients in CRGs-cluster B and gene-cluster A had significantly higher risk scores than the other clusters (Fig. 6E and F). Subsequently, we investigated the relevance of risk score and OS. No matter in the entire cohort, training cohort or validation cohort, patients with a higher risk score had significantly shorter OS (Fig. 7A). The time-dependent ROC curve also reflected the good prediction ability of this score model, while the AUC value of the entire cohort was 0.607, 0.646 and 0.646 for 1 year, 3 years and 5 years, respectively (Fig. 7B). Finally, the survival outcome, risk status and expression level of six genes were shown in Fig. 7C-E. Besides, nomogram model based on OS and other clinical data also showed better prediction ability of the observed OS at 1, 3 and 5 years of survival (Fig. 7F and G). Tumor-infiltrating immune cell analysis also revealed that risk score was positively correlated with gamma-delta T cell (γδ T cell), plasma cell, natural killer cell, neutrophil, monocyte and macrophage M1/M2, and negatively correlated with Treg cell, monocyte and macrophage M0 (Fig. 8A and S2A).

Prediction Of Reaction To Chemotherapy In Oc

On account of the close association between higher risk score and worse prognosis, we then explored the predictability of the score model in guiding OC medical treatment. “pRRophetic” package in R was applied to predict the half maximal inhibitory concentration (IC50) of frequent chemotherapeutic drugs or signaling pathway inhibitors against tumors in OC patients. We identified 77 drugs with significant differences in IC50 between high-risk and low-risk groups. As shown in Fig. 8B, several commonly used chemotherapeutic drugs against OC in clinical, including Bleomycin, Cisplatin, Cytarabine, Docetaxel, Doxorubicin, Etoposide, Gemcitabine, Mitomycin C, Pazopanib, Sorafenib and Vinorelbine, had significantly lower IC50 in higher risk score group, indicating that these patients were more probably sensitive to those drugs. However, ABT-888 showed the opposite tendency.

Ovarian cancer is the most lethal gynecological tumor. On average, 46% of OC patients can survive over 5 years after diagnosis, while late-stage OC has a 5-year survival rate of 29%, in contrast to 92% of early stage(22). Unlike other gynecological malignant tumors, which can have obvious symptoms such as abnormal vaginal bleeding or discharge, early-stage OC usually does not represent any specific discomfort or clinical symptom since the ovaries hidden deep in the abdominal cavity. According to this, 75% of OC patients are diagnosed in late-stage (FIGO stage III/IV), and over 70% of patients suffer tumor recurrence after surgery or chemotherapy(23). Nowadays, the commonly-used treatment for OC is comprehensive staging operation followed by repeated platinum-based chemotherapies and long-term monitoring of biological markers, mostly CA125(24). However, as the most widely used biomarker of OC, CA125 levels are only increased in only 50% of FIGO stage I OC patients, and CA125 seems not so reliable with a relatively lower specificity and sensitivity for screening and diagnosing OC(25, 26). Recently, newly identified BRCA gene mutations and homologous recombination deficiency (HRD) have shown great potential in prediction of OC(27, 28). Besides, newly-discovered molecular biomarkers especially multiple-gene-based risk models often show better prediction ability in the prognosis of OC patients.

With the development of high-throughput sequencing technology and bioinformatics analysis, many large-scale public databases such as TCGA(29) and GEO(30)have revealed the genome profiles of many cancers. Access and analysis of these databases provide us with an overview of the genetic landscape and help us seek new biomarkers and novel therapeutic strategies as well as predict the prognosis of these cancers(31–33). For ovarian cancer, for example, Zhang et al.(34) utilized TCGA database to construct a prognosis model based on 17 m6a-related genes, while Ye et al.(35) combined TCGA and GEO database and defined a risk signature of ferroptosis-related genes. Both studies presented excellent predictive ability in OC patients’ survival time.

As one of the most important trace elements in the human body, copper functions in many important biological processes such as energy generation, iron acquisition, oxygen transport, cellular metabolism, blood coagulation, signal transduction, and so on(36, 37). Besides, previous studies also demonstrated that copper was significantly increased in the circulation system and ascites fluid of OC patients(38, 39), and dysregulation of ATP7A/ATP7B was correlated with platinum resistance in OC(40). Cuproptosis, a newly recognized RCD type by Tsvetkov et al.(17), discloses the important role of copper in cell death. However, only a few studies have focused on cuproptosis and cancer. To date, our research is the first to explore the relationship between cuproptosis and OC. In our research, 19 CRGs were obtained from previous studies and were considered closely related to copper metabolism and cuproptosis. Surprisingly, compared with 88 normal ovarian tissues, all 19 CRGs represented a significant difference in 379 OC patients, and most of them showed an up-regulate tendency. This indicated the prevalence of cuproptosis in OC patients. Besides, combining analysis of OS in TCGA and GEO databases also demonstrated that 12 CRGs had a great impact on the survival time of OC patients, which also proved the importance of cuproptosis in OC.

In order to better understand the important role of cuproptosis in the occurrence and progression of OC, we discovered the effect of some crucial CRGs knockdown with siRNAs in OC cell line. FDX1 was demonstrated to be the key regulator of copper-induced cell death by Tsvetkov et al.(17), and the deletion of FDX1 conferred resistance to cuproptosis in lung cancer cell line. In our study, FDX1 knockdown significantly enhanced the proliferation and migration abilities of OC cells, which was consistent with previous research and our expectations. Besides, combining analysis of OS also showed that OC patients with higher FDX1 expression had obviously better prognosis (Fig. 2A), which also supported our experiment result that FDX1 low expression correlated with a higher potential for malignancy in OC. In our study, knockdown of another cuproptosis-related core gene, LIAS, the key enzyme of lipoic acid biosynthesis, showed different effects in OC cells. A pan-cancers bioinformatic analysis of LIAS (41) suggested that high LIAS expression is correlated with good OS, progression-free survival, and post-progression survival in OC patients. However, our study found that knockdown of LIAS could obviously damage the proliferation and migration abilities of OC cells, which might be because of the important role of LIAS in protein lipoylation and activation of key kinases in the TCA cycle. The potential role of LIAS and cuproptosis in OC needs further exploration. What’s more, we also noticed that knockdown of SLC31A1 could severely suppress the malignant potential of OC cells, even the ability for cells to form colonies had been completely eliminated. As one of the key elements of cuproptosis, SLC31A1 is predominantly localized to the plasma membrane and is responsible for copper uptake(42). Our research first emphasized the importance of SLC31A1 and cuproptosis in OC, so as to provide new potential target for OC treatment.

To better understand the relationship between cuproptosis and OC, we next divided all 791 patients into two subgroups according to the expression of CRGs and we discovered the OC patients with relatively higher CRGs expression showed shorter OS than those with lower CRGs expression. GSVA analysis also found some functional pathways notably distinct between two subgroups, including pathways of translation initiation, RNA polymerase initiation, ubiquitin ligase complex, nucleotide excision repair, mismatch repair, telomere maintenance, JAK-STAT signaling pathway, spliceosome and so on, which were demonstrated to be closely related to OC by previous studies(43–52). What’s more, the immune cell infiltration model predicted that B cell, CD8 positive T cell, eosinophil, macrophage, MDSC, Natural Killer cell and Natural Killer T cell were significantly lower-expressed in the subgroup with high CRGs expression, suggesting that the effect of immunotherapy might be affected by the level of cuproptosis in OC patients.

In order to better understand the important role of cuproptosis in OC, 70 differentially-expressed genes were identified among two subgroups. GO analysis revealed that those genes were enriched in the pathway of acetyl − CoA biosynthetic process, which was consistent with previous study(17). To better verify the correlation between those genes and OC, we again divided all 791 OC patients into two subgroups according to the expression of those differentially-expressed genes. As we expected, the subgroup of patients with higher expression of those genes had a shorter survival time, showed the same tendency with two CRG subgroups. These results once again illustrated the importance of cuproptosis in OC.

Recently, the identification of novel biomarkers has become a new direction in tumor treatment. Good biomarkers should be able to distinguish patients with different risks and make more accurate predictions on the survival time of patients, so as to provide better and individual treatment for every patient. Considering the important role of cuproptosis in OC, our research subsequently performed LASSO Cox regression and multivariate COX regression and 6 genes (BMI1, NRIP1, PSAT1, MAD2L1, SCG5, CXCL14) were selected to construct a prognosis risk score model from those 70 genes. In the training cohort, higher risk score of OC patients was correlated with poor survival, as well as in the validation cohort, which reflected the good prediction ability of this prognostic risk score model. And then, we built a predictive nomogram by integrating the risk score and age of patients, thus the reliability and prediction ability of this risk model were fully increased.

It is well known that the tumor immune microenvironment is closely related to most of the biological processes of cancer, including tumorigenesis, progression, invasion and metastasis, even the response to chemotherapy or radiotherapy treatment. Several types of immune cells have also been shown to play a special role in the tumor microenvironment(53, 54). For example, γδ T cells can promote the anti-tumor function of adaptive immune cells, but conversely also show tumor-prompting effects in some kinds of cancer(55, 56). Chen et al.(57) discovered that the percentage of γδ T cells in OC tissues was greater than in benign ovarian tumors and normal ovarian tissues. Besides, a previous study by Lee et al.(58) demonstrated that neutrophils could facilitate OC cells’ metastasizing from primary site to omentum. Furthermore, macrophage-polarized can represent tumorigenesis effect since in most malignant tumor including OC, and high density of M2-polarized macrophage correlated with poor prognosis in OC(59, 60). In our research, the percentage of M2-polarized macrophages, γδ T cells and neutrophils in tumor were all positively correlated with risk score, which is consistent with previous studies since in our risk model, higher risk score was closely related to poor prognosis. Meanwhile, the percentage of several important immune cells was also strongly linked to risk score in our research. As Aran et al.(61) believed, the inflammatory reaction caused by some kinds of infiltrating immune cells could lead to genetic mutation of tumor cells, so as to change the prognosis of patients. Similarly, we also considered that the poor prognosis in patients with a higher risk score might be closely related to the difference in infiltrating immune cells. Until now, immunotherapy has been the most promising new therapy for malignant tumors, and it is important to distinguish patients suitable for immunotherapy in clinical practice. A reliable prediction model can provide strong support for treatment.

As one of the most significant treatments for malignant tumors, chemotherapy plays an important role in preoperative and postoperative adjuvant treatment, palliative treatment, lightening tumor load, preventing recurrence, prolonging survival time and even achieving complete cure of cancer. Especially in OC patients, the current standard treatment is primary aggressive cytoreductive surgery followed by platinum-based chemotherapy(62). Besides, chemotherapy is basically the only effective treatment that can alleviate recurrent OC. However, some platinum-refractory patients show obvious resistance to drugs at the first treatment and even progress during chemotherapy(63). Although Poly ADP Ribose Polymerase (PARP) inhibitors were widely used recently and showed a good effect in improving the survival time of OC patients, inevitable drug resistance still occurred frequently(64). Therefore, predicting the sensitivity before using chemotherapy drugs has become a new research interest. With the progression of mechanisms in drug resistance and the widespread usage of high-throughput sequencing technology and bioinformatics analysis, it has become possible to predict the response of chemotherapy in cancers. In our study, 77 drugs including commonly used chemotherapeutic drugs, cellular signaling pathway activators and inhibitors were considered to have possibly different chemotherapeutic response between higher and lower risk score groups. It is remarkable that some famous, frequently-used drugs in first-line or second-line chemotherapy of OC, such as Bleomycin, Cisplatin, Cytarabine, Docetaxel, Doxorubicin, Etoposide, Gemcitabine, Mitomycin C, Sorafenib and Vinorelbine, have a significantly lower IC50 value in the higher risk score group, suggesting that those patients may be more sensitive to these drugs above(63). Besides, patients with a higher risk score also showed more sensitivity to Pazopanib, a kind of angiogenesis inhibitor suitable for late-stage OC(65), and this can provide evidence for clinical practice. To our surprise, ABT-888, also known as Veliparib, a widely-used PARP inhibitor in OC patients(66), showed the opposite effect in higher risk score group with a higher IC50 value. This suggests that usage of this drug in patients with a higher risk score requires more caution and further investigation needs to be done. However, with the decreasing cost of high-throughput sequencing technology, it is necessary and beneficial for every OC patient to perform sequencing after diagnosis, which can help to predict drug sensitivity based on the risk score model before initiating treatment, so as to provide guidance on drug selection and achieve the best therapeutic effect.

Notably, this study still had many limitations. First, no external database was used to validate the model. Second, in order to build a larger database containing survival time as possible, GSE13876 dataset was adopted, resulting in the deficiency of partial pathology information. Third, given that the prognostic signature and risk score model was built and validated by public databases, further experimental verification is still needed.

In conclusion, our study suggests that cuproptosis may play an important role in the progression of OC. We combinedly analyze cuproptosis-related differentially-associated genes in TCGA and GEO database and construct a risk model with OS prognosis, and preliminarily reveal the close relationship between cuproptosis and tumor-infiltrating immune cells and chemotherapy resistance, which may provide new understandings to the treatment of OC.

Data acquisition

RNA transcriptome profiling data in Fragments Per Kilobase of exon model per Million mapped fragments (FPKM) form and clinical data of 379 OC patients’ tissues from the Cancer Genome Atlas (TCGA) dataset and FPKM form of 88 normal ovarian tissues from the Genotype-Tissue Expression (GTEx) dataset were downloaded from the UCSC Xena project (http://xena.ucsc.edu/)(67). Somatic mutation and copy number variation (CNV) data of OC patients in TCGA dataset were also downloaded from UCSC Xena project. MuTech2 algorithm(68) was applied to calculate the mutation of CRGs and “maftools” package(69) in R helped mapping. CNV values were reorganized and visualized by R. Microarray data based on GPL7759 platform and clinicopathological data of 415 OC patients in the GSE13876 dataset(70) were obtained from Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/)(30). Expression of CRGs was extracted from the datasets above by Perl software and “Limma” package in R(71) was used to compare CRGs between different groups.

Bioinformatic Analysis

“ConsensusClusterPlus” package(72) was used to divide OC patients into different clusters. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were conducted by “clusterProfiler” package(73) in R, while Gene Set Variation Analysis (GSVA) and Gene Set Enrichment Analysis (GSEA) was ran to ascertain the different pathways between two clusters using “GSVA”(74) and “GSEABase”(75) packages in R.

Cell-Type Identification by Estimating Relative Subsets of RNA Transcripts (CIBERSORT) algorithm(76) in R was adopted to identify the ratio of tumor-infiltrating immune cell in all OC patients. Correlation between risk score and infiltrating score of each kind of immune cells was calculated and visualized by scatter diagrams. A heatmap showing gene expression differences was performed in “pheatmap” package in R. The Kaplan-Meier curve was calculated by “survminer” package in R and used to compared the OS of patients with OC.

Construction Of A Prognostic Signature

We randomly allocated 791 OC patients into training group (n = 396) and validation group (n = 395) in a ratio of 1:1. LASSO Cox regression was performed by “glmnet” package(77) in R and 10 genes were selected from 70 differentially-expressed genes. Multivariate Cox regression was subsequently conducted by “survival” package in R to construct a risk model for each individual as a prognostic signature. The risk score was calculated by the sum of each gene expression level multiple its coefficients. The median value of the risk score was defined as the cut-off value of high-risk group and low-risk group in both training cohort, validation cohort and entire cohort. Overall survival (OS) was compared among different risk groups by Kaplan–Meier analysis in all three cohorts. The time-dependent ROC curve was drawn by “timeROC” package(78) in R.

A nomogram containing risk score and age was performed by “rms” package in R. The accuracy of nomogram was then verified by time-dependent calibration curves and a multivariate Cox regression analysis was used to confirm the independence of risk factor as an OS prediction in OC patients.

For predicting clinical chemotherapeutic response, “pRRophetic” package(79) in R was utilized to prognose the half maximal inhibitory concentration (IC50) of every commonly used chemotherapy drugs in all OC patients. “Limma” package was subsequently used to compare the IC50 between high- and low-risk groups.

Cell Experiments

The human OC cell line SK-OV-3 was purchased from ATCC and cultured in complete McCoy’s 5A medium (Gibco, USA) with 10% FBS (Gibco, USA). Cell line was cultured in a humidified atmosphere with 5% CO2 at 37℃. siRNAs designed and synthesized by GenePharma (Suzhou, China) were transfected into SK-OV-3 cell line using Lipofectamine RNAiMAX (Invitrogen, USA) according to the protocol provided by the manufacturer. The sequence of siRNA-FDX1 was sense strand 5’-GUGAUUCUCUGCUAGAUGUTT-3’ and antisense strand 5’-ACAUCUAGCAGAGAAUCACTT-3’. The sequence of siRNA-LIAS was sense strand 5’-GCACCUGGGAUGAAUAUAATT-3’ and antisense strand 5’-UUAUAUUCAUCCCAGGUGCTT-3’. The sequence of siRNA-SLC31A1 was sense strand 5’-GAGAGAGCCUGCUGCGUAATT-3’ and antisense strand 5’- UUACGCAGCAGGCUCUCUCTT-3’. The sequence of negative control siRNA (siRNA-NC) was sense strand 5’-UUCUCCGAACGUGUCACGUTT-3’ and antisense strand 5’-ACGUGACACGUUCGGAGAATT-3’.

For RNA extraction and quantitative real-time PCR, total RNA was extracted using SteadyPure Universal RNA Extraction Kit (AG21017, Accurate Biotechnology, China) and reverse transcripted to cDNA using Evo M-MLV RT Premix for qPCR (Accurate Biotechnology, China). Quantitative real-time PCR (qRT-PCR) was performed by CFX96 real-time PCR detection system (Bio-Rad, USA) using SYBR Green Premix Pro Taq HS qPCR Kit (Accurate Biotechnology, China). GAPDH was utilized as an internal standard control. The gene-specific primers were synthesized by GENEray Biotechnology (Shanghai, China) as follows: FDX1 forward 5’-CCTGGCTTGTTCAACCTGTCA-3’, reverse 5’-CCAACCGTGATCTGTCTGTTAGTC-3’; LIAS forward 5’-GTATGTGAGGAAGCTCGATGTC-3’, reverse 5’-CACCCATCAACATGATCGTGG-3’; SLC31A1 forward 5’-GGGGATGAGCTATATGGACTCC-3’, reverse 5’-TCACCAAACCGGAAAACAGTAG-3’; GAPDH forward 5’-ATCACCATCTTCCAGGAGCGA-3’, reverse 5’- CCTTCTCCATGGTGGTGAAGAC-3’. Relative RNA expression was calculated by the relative quantification method (2-ΔΔCT).

For CCK-8 assay, 2×10³ cells were evenly seeded into 96-well plates for various time periods. Cell Counting Kit-8 kit (Beyotime, China) was then used to measure cell proliferation ability in accordance with the manufacturer’s instructions. Optical Density at 450nm (OD450) was measured by fluorescence microplate reader after incubating at 37℃ for 1 hour. The moment of cell transfection was defined as 0 hours and CCK-8 assay was performed every 24 hours.

For colony formation assay, 24 hours after cell transfection, SK-OV-3 cells were digested and replanted into 6-well plates with 2×10³ cells per well. After 7 days of culture, the cells were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. Each well was photographed separately.

We used a 24-well plate Transwell system (8.0µm pore size, Falcon) to determine cell migration. About 24 hours after cell transfection, 2×10⁴ cells were seeded into the upper chamber and incubated in a 10% FBS complete culture medium for 24 hours. Next, the cells were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet, and subsequently photographed under microscope. For the wound healing assay, culture inserts (ibidi, German) were utilized to create scratches among cells. 24 hours after cell transfection, 2×10⁴ cells were seeded into every quadrant of inserts in complete culture medium. Inserts were removed after 12 hours and cells were then cultured in FBS-free medium. Photographing under microscope was processed at 0 hours and 8 hours after removal for comparing wound healing abilities.

Statistical analysis

An independent T-test and one-way ANOVA were adopted to make comparison between different groups, while the Pearson correlation coefficient was employed to reflect correlations between CRGs. We considered a level of p-value less than 0.05 to be statistically significant (*: p-value < 0.05, **: p-value < 0.01, ***: p-value < 0.001). Perl version 5.30.0 was used to reorganize the information into a matrix. All the analyses above were performed by R version 4.1.3 and GraphPad Prism 8.0.1.

OC, ovarian cancer; OS, overall survival; TCGA, the Cancer Genome Atlas; GEO, Gene Expression Omnibus; IC50, half maximal inhibitory concentration; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; GSVA, Gene Set Variation Analysis; GESA, Gene Set Enrichment Analysis; FIGO, the International Federation of Gynecology and Obstetrics; CRG, cuproptosis-related gene; RCD, regulated cell death; TCA, tricarboxylic acid; FPKM, Fragments Per Kilobase of exon model per Million mapped fragments; CNV, copy number variation; GTEx, Genotype-Tissue Expression; CIBERSORT, Cell-Type Identification by Estimating Relative Subsets of RNA Transcripts; CA125, carbohydrate antigen 125; HRD, homologous recombination deficiency; PARP, Poly ADP Ribose Polymerase.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

The datasets generated and analysed during the current study are available in the TCGA repository, http://xena.ucsc.edu, and GEO repository, https://www.ncbi.nlm.nih.gov/geo (GSE13876).

Competing interests

The authors declare that they have no competing interests.

Funding

This work was supported by the National Key R&D Program of China (2022YFC2704200) to Shuzhong Yao, National Natural Science Foundation of China (No. 81874102, 82072874 to Shuzhong Yao; No. 82072884, 81872128, 81502226 to Junxiu Liu), China Postdoctoral Science Foundation (No. 2022M713591 to Yili Chen), Guangzhou Science and Technology Program (No. 202002020043 to Shuzhong Yao) and Sun Yat-sen University Clinical Research Foundation of 5010 Project (No. 2017006 to Shuzhong Yao).

Authors contributions

QZ designed the study and processed bioinformatics analysis under the guidance of SY and JL. SW and YC completed cell experiments. QD participated in data analysis and language editing. CZ helped with data acquisition. SY and JL contributed to the revision of the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We thank the study participants for their contributions to this study. We also acknowledge TCGA, GEO, GTEx database, UCSC Xena project and the developers of all the R packages mentioned in our study.

Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA: a cancer journal for clinicians. 2021;71(3):209-49.
Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA: a cancer journal for clinicians. 2018;68(1):7-30.
Lee JY, Kim S, Kim YT, Lim MC, Lee B, Jung KW, et al. Changes in ovarian cancer survival during the 20 years before the era of targeted therapy. BMC cancer. 2018;18(1):601.
Kim BE, Nevitt T, Thiele DJ. Mechanisms for copper acquisition, distribution and regulation. Nature chemical biology. 2008;4(3):176-85.
Bandmann O, Weiss KH, Kaler SG. Wilson's disease and other neurological copper disorders. The Lancet Neurology. 2015;14(1):103-13.
Tümer Z, Møller LB. Menkes disease. European journal of human genetics : EJHG. 2010;18(5):511-8.
Ding X, Jiang M, Jing H, Sheng W, Wang X, Han J, et al. Analysis of serum levels of 15 trace elements in breast cancer patients in Shandong, China. Environmental science and pollution research international. 2015;22(10):7930-5.
Zhang X, Yang Q. Association between serum copper levels and lung cancer risk: A meta-analysis. The Journal of international medical research. 2018;46(12):4863-73.
Baltaci AK, Dundar TK, Aksoy F, Mogulkoc R. Changes in the Serum Levels of Trace Elements Before and After the Operation in Thyroid Cancer Patients. Biological trace element research. 2017;175(1):57-64.
Aubert L, Nandagopal N, Steinhart Z, Lavoie G, Nourreddine S, Berman J, et al. Copper bioavailability is a KRAS-specific vulnerability in colorectal cancer. Nature communications. 2020;11(1):3701.
Chen F, Wang J, Chen J, Yan L, Hu Z, Wu J, et al. Serum copper and zinc levels and the risk of oral cancer: A new insight based on large-scale case-control study. Oral diseases. 2019;25(1):80-6.
Pavithra V, Sathisha TG, Kasturi K, Mallika DS, Amos SJ, Ragunatha S. Serum levels of metal ions in female patients with breast cancer. Journal of clinical and diagnostic research : JCDR. 2015;9(1):BC25-c7.
Denoyer D, Clatworthy SAS, Cater MA. Copper Complexes in Cancer Therapy. Metal ions in life sciences. 2018;18.
Babak MV, Ahn D. Modulation of Intracellular Copper Levels as the Mechanism of Action of Anticancer Copper Complexes: Clinical Relevance. Biomedicines. 2021;9(8).
Tang D, Kang R, Berghe TV, Vandenabeele P, Kroemer G. The molecular machinery of regulated cell death. Cell research. 2019;29(5):347-64.
Koren E, Fuchs Y. Modes of Regulated Cell Death in Cancer. Cancer discovery. 2021;11(2):245-65.
Tsvetkov P, Coy S, Petrova B, Dreishpoon M, Verma A, Abdusamad M, et al. Copper induces cell death by targeting lipoylated TCA cycle proteins. Science (New York, NY). 2022;375(6586):1254-61.
Zhang G, Sun J, Zhang X. A novel Cuproptosis-related LncRNA signature to predict prognosis in hepatocellular carcinoma. Scientific reports. 2022;12(1):11325.
Han J, Hu Y, Liu S, Jiang J, Wang H. A Newly Established Cuproptosis-Associated Long Non-Coding RNA Signature for Predicting Prognosis and Indicating Immune Microenvironment Features in Soft Tissue Sarcoma. Journal of oncology. 2022;2022:8489387.
Lv H, Liu X, Zeng X, Liu Y, Zhang C, Zhang Q, et al. Comprehensive Analysis of Cuproptosis-Related Genes in Immune Infiltration and Prognosis in Melanoma. Frontiers in pharmacology. 2022;13:930041.
Bian Z, Fan R, Xie L. A Novel Cuproptosis-Related Prognostic Gene Signature and Validation of Differential Expression in Clear Cell Renal Cell Carcinoma. Genes. 2022;13(5).
Lheureux S, Gourley C, Vergote I, Oza AM. Epithelial ovarian cancer. Lancet (London, England). 2019;393(10177):1240-53.
Matulonis UA, Sood AK, Fallowfield L, Howitt BE, Sehouli J, Karlan BY. Ovarian cancer. Nature reviews Disease primers. 2016;2:16061.
Kuroki L, Guntupalli SR. Treatment of epithelial ovarian cancer. BMJ (Clinical research ed). 2020;371:m3773.
Dochez V, Caillon H, Vaucel E, Dimet J, Winer N, Ducarme G. Biomarkers and algorithms for diagnosis of ovarian cancer: CA125, HE4, RMI and ROMA, a review. Journal of ovarian research. 2019;12(1):28.
Kobayashi H, Yamada Y, Sado T, Sakata M, Yoshida S, Kawaguchi R, et al. A randomized study of screening for ovarian cancer: a multicenter study in Japan. International journal of gynecological cancer : official journal of the International Gynecological Cancer Society. 2008;18(3):414-20.
Moschetta M, George A, Kaye SB, Banerjee S. BRCA somatic mutations and epigenetic BRCA modifications in serous ovarian cancer. Annals of oncology : official journal of the European Society for Medical Oncology. 2016;27(8):1449-55.
Ledermann JA, Drew Y, Kristeleit RS. Homologous recombination deficiency and ovarian cancer. European journal of cancer (Oxford, England : 1990). 2016;60:49-58.
Tomczak K, Czerwińska P, Wiznerowicz M. The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge. Contemporary oncology (Poznan, Poland). 2015;19(1a):A68-77.
Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic acids research. 2002;30(1):207-10.
Zheng Q, Min S, Zhou Q. Identification of potential diagnostic and prognostic biomarkers for LUAD based on TCGA and GEO databases. Bioscience reports. 2021;41(6).
Fakih M, Ouyang C, Wang C, Tu TY, Gozo MC, Cho M, et al. Immune overdrive signature in colorectal tumor subset predicts poor clinical outcome. The Journal of clinical investigation. 2019;129(10):4464-76.
Huang R, Mao M, Lu Y, Yu Q, Liao L. A novel immune-related genes prognosis biomarker for melanoma: associated with tumor microenvironment. Aging. 2020;12(8):6966-80.
Zhang C, Liu J, Guo H, Hong D, Ji J, Zhang Q, et al. m6A RNA methylation regulators were associated with the malignancy and prognosis of ovarian cancer. Bioengineered. 2021;12(1):3159-76.
Ye Y, Dai Q, Li S, He J, Qi H. A Novel Defined Risk Signature of the Ferroptosis-Related Genes for Predicting the Prognosis of Ovarian Cancer. Frontiers in molecular biosciences. 2021;8:645845.
Ruiz LM, Libedinsky A, Elorza AA. Role of Copper on Mitochondrial Function and Metabolism. Frontiers in molecular biosciences. 2021;8:711227.
Ge EJ, Bush AI, Casini A, Cobine PA, Cross JR, DeNicola GM, et al. Connecting copper and cancer: from transition metal signalling to metalloplasia. Nature reviews Cancer. 2022;22(2):102-13.
Onuma T, Mizutani T, Fujita Y, Yamada S, Yoshida Y. Copper content in ascitic fluid is associated with angiogenesis and progression in ovarian cancer. Journal of trace elements in medicine and biology : organ of the Society for Minerals and Trace Elements (GMS). 2021;68:126865.
Lin S, Yang H. Ovarian cancer risk according to circulating zinc and copper concentrations: A meta-analysis and Mendelian randomization study. Clinical nutrition (Edinburgh, Scotland). 2021;40(4):2464-8.
Lukanović D, Herzog M, Kobal B, Černe K. The contribution of copper efflux transporters ATP7A and ATP7B to chemoresistance and personalized medicine in ovarian cancer. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie. 2020;129:110401.
Cai Y, He Q, Liu W, Liang Q, Peng B, Li J, et al. Comprehensive analysis of the potential cuproptosis-related biomarker LIAS that regulates prognosis and immunotherapy of pan-cancers. Frontiers in oncology. 2022;12:952129.
Schweigel-Röntgen M. The families of zinc (SLC30 and SLC39) and copper (SLC31) transporters. Current topics in membranes. 2014;73:321-55.
Crosbie EJ, Ryan NAJ, McVey RJ, Lalloo F, Bowers N, Green K, et al. Assessment of mismatch repair deficiency in ovarian cancer. Journal of medical genetics. 2021;58(10):687-91.
Hu X, Meng Y, Xu L, Qiu L, Wei M, Su D, et al. Cul4 E3 ubiquitin ligase regulates ovarian cancer drug resistance by targeting the antiapoptotic protein BIRC3. Cell death & disease. 2019;10(2):104.
Pal T, Akbari MR, Sun P, Lee JH, Fulp J, Thompson Z, et al. Frequency of mutations in mismatch repair genes in a population-based study of women with ovarian cancer. British journal of cancer. 2012;107(10):1783-90.
Liu T, Wei Q, Jin J, Luo Q, Liu Y, Yang Y, et al. The m6A reader YTHDF1 promotes ovarian cancer progression via augmenting EIF3C translation. Nucleic acids research. 2020;48(7):3816-31.
Saldivar JS, Wu X, Follen M, Gershenson D. Nucleotide excision repair pathway review I: implications in ovarian cancer and platinum sensitivity. Gynecologic oncology. 2007;107(1 Suppl 1):S56-71.
Li Y, Guo H, Jin C, Qiu C, Gao M, Zhang L, et al. Spliceosome-associated factor CTNNBL1 promotes proliferation and invasion in ovarian cancer. Experimental cell research. 2017;357(1):124-34.
Cornelison R, Dobbin ZC, Katre AA, Jeong DH, Zhang Y, Chen D, et al. Targeting RNA-Polymerase I in Both Chemosensitive and Chemoresistant Populations in Epithelial Ovarian Cancer. Clinical cancer research : an official journal of the American Association for Cancer Research. 2017;23(21):6529-40.
Song M, Yeku OO, Rafiq S, Purdon T, Dong X, Zhu L, et al. Tumor derived UBR5 promotes ovarian cancer growth and metastasis through inducing immunosuppressive macrophages. Nature communications. 2020;11(1):6298.
Sun Y, Tao W, Huang M, Wu X, Gu J. Genetic variants in telomere-maintenance genes are associated with ovarian cancer risk and outcome. Journal of cellular and molecular medicine. 2017;21(3):510-8.
Bagratuni T, Mavrianou N, Gavalas NG, Tzannis K, Arapinis C, Liontos M, et al. JQ1 inhibits tumour growth in combination with cisplatin and suppresses JAK/STAT signalling pathway in ovarian cancer. European journal of cancer (Oxford, England : 1990). 2020;126:125-35.
Pitt JM, Marabelle A, Eggermont A, Soria JC, Kroemer G, Zitvogel L. Targeting the tumor microenvironment: removing obstruction to anticancer immune responses and immunotherapy. Annals of oncology : official journal of the European Society for Medical Oncology. 2016;27(8):1482-92.
Hinshaw DC, Shevde LA. The Tumor Microenvironment Innately Modulates Cancer Progression. Cancer research. 2019;79(18):4557-66.
Li Y, Li G, Zhang J, Wu X, Chen X. The Dual Roles of Human γδ T Cells: Anti-Tumor or Tumor-Promoting. Frontiers in immunology. 2020;11:619954.
Silva-Santos B, Serre K, Norell H. γδ T cells in cancer. Nature reviews Immunology. 2015;15(11):683-91.
Chen X, Shang W, Xu R, Wu M, Zhang X, Huang P, et al. Distribution and functions of γδ T cells infiltrated in the ovarian cancer microenvironment. Journal of translational medicine. 2019;17(1):144.
Lee W, Ko SY, Mohamed MS, Kenny HA, Lengyel E, Naora H. Neutrophils facilitate ovarian cancer premetastatic niche formation in the omentum. The Journal of experimental medicine. 2019;216(1):176-94.
Allavena P, Sica A, Solinas G, Porta C, Mantovani A. The inflammatory micro-environment in tumor progression: the role of tumor-associated macrophages. Critical reviews in oncology/hematology. 2008;66(1):1-9.
Baci D, Bosi A, Gallazzi M, Rizzi M, Noonan DM, Poggi A, et al. The Ovarian Cancer Tumor Immune Microenvironment (TIME) as Target for Therapy: A Focus on Innate Immunity Cells as Therapeutic Effectors. International journal of molecular sciences. 2020;21(9).
Aran D, Sirota M, Butte AJ. Systematic pan-cancer analysis of tumour purity. Nature communications. 2015;6:8971.
Orr B, Edwards RP. Diagnosis and Treatment of Ovarian Cancer. Hematology/oncology clinics of North America. 2018;32(6):943-64.
Freimund AE, Beach JA, Christie EL, Bowtell DDL. Mechanisms of Drug Resistance in High-Grade Serous Ovarian Cancer. Hematology/oncology clinics of North America. 2018;32(6):983-96.
D'Andrea AD. Mechanisms of PARP inhibitor sensitivity and resistance. DNA repair. 2018;71:172-6.
McLachlan J, Banerjee S. Pazopanib in ovarian cancer. Expert review of anticancer therapy. 2015;15(9):995-1005.
Coleman RL, Fleming GF, Brady MF, Swisher EM, Steffensen KD, Friedlander M, et al. Veliparib with First-Line Chemotherapy and as Maintenance Therapy in Ovarian Cancer. The New England journal of medicine. 2019;381(25):2403-15.
Goldman MJ, Craft B, Hastie M, Repečka K, McDade F, Kamath A, et al. Visualizing and interpreting cancer genomics data via the Xena platform. Nature biotechnology. 2020;38(6):675-8.
Cibulskis K, Lawrence MS, Carter SL, Sivachenko A, Jaffe D, Sougnez C, et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nature biotechnology. 2013;31(3):213-9.
Mayakonda A, Lin DC, Assenov Y, Plass C, Koeffler HP. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome research. 2018;28(11):1747-56.
Crijns AP, Fehrmann RS, de Jong S, Gerbens F, Meersma GJ, Klip HG, et al. Survival-related profile, pathways, and transcription factors in ovarian cancer. PLoS medicine. 2009;6(2):e24.
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic acids research. 2015;43(7):e47.
Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics (Oxford, England). 2010;26(12):1572-3.
Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics : a journal of integrative biology. 2012;16(5):284-7.
Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC bioinformatics. 2013;14:7.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(43):15545-50.
Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, et al. Robust enumeration of cell subsets from tissue expression profiles. Nature methods. 2015;12(5):453-7.
Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. Journal of statistical software. 2010;33(1):1-22.
Blanche P, Dartigues JF, Jacqmin-Gadda H. Estimating and comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Statistics in medicine. 2013;32(30):5381-97.
Geeleher P, Cox N, Huang RS. pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels. PloS one. 2014;9(9):e107468.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

A cuproptosis-related gene cluster in prediction of ovarian cancer prognosis and chemotherapeutic response

Status:

Version 1

Abstract

Background

Results

Conclusion

Figures

Background

Result

Expression of Cuproptosis-related genes in ovarian cancer

Crgs Correlate With Oc Proliferation And Migration In Vitro

Consensus Cluster Analysis Of Crgs Identified Two Oc Subgroups

Genes Expressed Differently Between Two Crg-related Subgroups

Construction Of A Crgs-related Prognostic Signature In Oc

Prediction Of Reaction To Chemotherapy In Oc

Discussion

Conclusion

Methods

Data acquisition

Bioinformatic Analysis

Construction Of A Prognostic Signature

Cell Experiments

Statistical analysis

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1